On 2020-05-11 01:41:57, user Sinai Immunol Review Project wrote:
Main findings<br />
The need for improved cellular profiling of host immune responses seen in COVID-19 has required the use of high-throughput technologies that can detail the immune landscape of these patients at high granularity. To fulfill that need, Chua et al. performed 3’ single-cell RNA sequencing (scRNAseq) on nasopharyngeal (or pooled nasopharyngeal/pharyngeal swabs) (NS), bronchiolar protected specimen brush (PSB), and broncheoalveolar lavage (BAL) samples from 14 COVID-19 patients with moderate (n=5) and critical (n=9, all admitted to the ICU; n=2 deaths) disease, according to WHO criteria. Four patients (n=2 with moderate COVID-19; n=2 with critical disease, n=1 on short-term non-invasive ventilation and n=1 on long-term invasive ventilation), were sampled longitudinally up to four times at various time points post symptom onset. In addition, multiple samples from all three respiratory sites (NS, PSB, BAL) were collected from two ICU patients on long-term mechanical ventilation, one of whom died a few days after the sampling procedure. Moreover, three SARS-CoV-2 negative controls, one patient diagnosed with Influenza B as well as two volunteers described as “supposedly healthy”, were included in this study with a total of n=17 donors and n=29 samples.
Clustering analysis of cells isolated from NS samples identified all major epithelial cell types, including basal, scretory, ciliated, and FOXN4+ cells as well as ionocytes; of particular note, a subset of basal cells was found to have a positive IFN? transcriptional signature, suggesting prior activation of these cells by the host immune system, likely in response to viral injury. In addition to airway epithelial cells, 6 immune cell types were identified and further subdivided into a total of 12 different subsets. These included macrophages (moMacs, nrMacs), DCs (moDCs, pDCs), mast cells, neutrophils, CD8 T (CTLs, lytic T cells), B, and NKT cells; however, seemingly neither NK nor CD4 T cells were detected and the Treg population lacked canonical expression of FoxP3, so it is unclear whether this population is truly represented.
Interestingly, secretory and ciliated cells in COVID-19 patients were shown to have upregulated ACE2 and coexpression with at least one S-priming protease indicative of viral infection; ACE2 expression on respiratory target cells increased by 2-3 fold in COVID-19 patients, compared to healthy controls. Notably, ciliated cells were mostly ACE2+/TMRPSS+, while secretory and FOXN4+ cells were predominantly ACE2+/TMRPSS+/FURIN+; accordingly, secretory and ciliated cells contained the highest number of SARS-CoV-2 infected cells. However, viral transcripts were generally low 10 days post symptom onset (as would be expected based on reduced viral shedding in later stages of COVID-19). Similarly, the authors report very low counts of immune cell-associated viral transcripts that are likely accounted for by the results of phagocytosis or surface binding. However, direct infection of macrophages by SARS-CoV-2 has previously been reported 1,2. Here, it is possible that these differences could be due to the different clinical stages and non-standardized gene annotation.
Pseudotime mapping of the obtained airway epithelial data suggested a direct differentiation trajectory from basal to ciliated cells (in contrast to the classical pathway from basal cells via secretory cells to terminally differentiated ciliated cells), driven by interferon stimulated genes (ISGs). Moreover, computational interaction analysis between these ACE2+ secretory/ciliated cells and CD8 CTLs indicated that upregulation of ACE2 receptor expression on airway epithelial cells might be induced by IFN?, derived from these lymphocytes. However, while IFN-mediated ACE2 upregulation in response to viral infections may generally be considered a protective component of the antiviral host response, the mechanism proposed here may be particularly harmful in the context of critical COVID-19, rendering these patients more susceptible to SARS-CoV-2 infection.
Moreover, direct comparisons between moderate and critical COVID-19 patient samples revealed fewer tissue-resident macs and monocyte-derived dendritic cells but increased frequencies of non-resident macs and neutrophils in critically ill COVID-19 patients. Notably, neutrophil infiltration in COVID-19 samples was significantly greater than in those obtained from healthy controls and the Influenza B patient. In addition, patients with moderate disease and those on short-term non-invasive ventilation had similar gene expression profiles (each n=1),; whereas, critical patients on long-term ventilation expressed substantially higher levels of pro-inflammatory and chemoattractant genes including TNF, IL1B, CXCL5, CCL2, and CCL3. However, no data on potentially decreasing gene expression levels related to convalescence were obtained. Generally, these profiles support findings of activated, inflammatory macrophages and CTLs with upregulated markers of cytotoxicity in critically ill COVID-19 patients. These inflammatory macrophages and CTLs may further contribute to pathology via apoptosis suggested by high CASP3 levels in airway epithelial cells. Interestingly, the CCL5/CCR5 axis was enriched among CTLs in PSB and BAL samples obtained from moderate COVID-19 patients; recently, a disruption of that axis using leronlimab was reported to induce restoration of the CD8 T cell count in critically ill COVID-19 patients 3.
Lastly, in critically ill COVID-19 patients, non-resident macrophages were found to have higher expression levels of genes involved in extravasation processes such as ITGAM, ITGAX and others. Conversely, endothelial cells were shown to express VEGFA and ICAM1, which are typical markers of macrophage/immune cell recruitment. This finding supports the notion that circulating inflammatory monocytes interact with dysfunctional endothelium to infiltrate damaged tissues. Of note, in the patient with influenza B, cellular patterns and expression levels of these extravasation markers were profoundly different from critically ill COVID-19.
Importantly, the aforementioned immune cell subsets were found equally in all three respiratory site samples obtained from two multiple-sample ICU donors, and there were no differences, with regards to upper vs. lower respiratory tract epithelial ACE2 expression. However, viral loads were higher in BAL samples as compared to NS samples, and lower respiratory tract macrophages showed overall greater pro-inflammatory potential, corresponding to higher CASP3 levels found in PSB and BAL samples. In general, the interactions between host airway epithelial and immune cells described in this preprint likely contribute to viral clearance in mild and moderate disease but might be excessive in critical cases and may therefore contribute to the observed COVID-19 immunopathology. Based on these findings and the discussed immune cell profiles above, the authors suggest the use of immunomodulatory therapies targeting chemokines and chemokine receptors, such as blockade of CCR1 by itself or in combination with CCR5, to treat COVID-19 associated hyperinflammation.
Limitations<br />
Technical<br />
In addition to the small sample size, it is unclear whether samples were collected at similar time points throughout the disease course of each patient, even with time since diagnosis normalized across patients. While sampling dates in relation to symptom onset are listed, it remains somewhat unclear what kind of samples were routinely obtained per patient at given time points (with the exception of the two patients with multiple sampling). Moreover, it would have been of particular interest (and technically feasible) to collect additional swabs from the convalescent ICU patient to generate a kinetic profile of chemokine gene expression levels, with respect to disease severity as well as onset of recovery. Again, with an n=1, the number of cases per longitudinal/multiple sampling subgroup is very limited, and, in addition to the variable sampling dates, overall time passed since symptom onset as well as disease symptoms and potential treatment (e.g. invasive vs non-invasive ventilation, ECMO therapy…) across all clinical subgroups, makes a comparative analysis rather difficult.
It is important to note that a lack of standardized gene annotation across different studies contributes to a significant degree of variability in characterizations of immune landscapes found in COVID-19 patients. As a result, inter-study comparisons are difficult to perform. For instance, an analysis of single-cell RNA sequencing performed on bronchoalveolar lavage samples by Bost et al. identified lymphoid populations that were not found in the present study. These include several enriched subtypes of CD4+ T cells and NK cells, among others. Ultimately, these transcriptomic descriptions will still need to be furthered with additional follow-up studies, including proteomic analysis, to move beyond speculation and towards substantive hypotheses.
Biological<br />
One additional limitation involved the use of the influenza B patient. Given that the patient suffered a rather mild form of the disease (no ICU admission or mechanical ventilation required, patient was discharged from hospital after 4 days) as opposed to the to authors’ assessment as a severe case, this patient may have served as an acceptable positive control for mild and some moderate COVID-19 patients. However, this approach should still be viewed cautiously, since the potential differences of pulmonary epithelial and immune cell pathologies induced by influenza compared to critical COVID-19 patients are still unclear. Moreover, it seems that one of the presumably healthy controls was recovering from a viral infection. Since it is unclear how a recent mild viral infection might have changed the respiratory cellular compartment and immune cell phenotype, this donor should have been excluded or not used as a healthy reference control.
Significance<br />
In general, this is a well-conducted study and provides a number of corroborative and interesting findings that contribute to our understanding of immune and non-immune cell heterogeneity in COVID-19 pathogenesis. Importantly, observations on ACE2 and ACE2 coexpression in airway epithelial cells generally corroborate previous reports. In addition, direct differentiation of IFN?+ basal cells to ACE2-expressing ciliated cells, as suggested by trajectory analysis, is a very interesting hypothesis, which, if confirmed, might contribute to progression of disease severity. The findings described in this preprint further suggest an important role for chemokines and chemokine receptors on immune cells, most notably macrophages and CTLs, which is highly relevant.
This review was undertaken by Matthew D. Park and Verena van der Heide as part of a project by students, postdocs and faculty at the Immunology Institute of the Icahn school of medicine, Mount Sinai.
References<br />
1. Chen, Y. et al. The Novel Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Directly Decimates Human Spleens and Lymph Nodes. Infectious Diseases (except HIV/AIDS) (2020) doi:10.1101/2020.03.27.20045427.<br />
2. Bost, P. et al. Host-viral infection maps reveal signatures of severe COVID-19 patients. Cell (2020) doi:10.1016/j.cell.2020.05.006.<br />
3. Patterson, B. K. et al. Disruption of the CCL5/RANTES-CCR5 Pathway Restores Immune Homeostasis and Reduces Plasma Viral Load in Critical COVID-19. medRxiv (2020).